Publisher’s version / Version de l'éditeur:
Journal of Analytical Atomic Spectrometry, 20, 11, pp. 1226-1231, 2005
READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE. https://nrc-publications.canada.ca/eng/copyright
Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca.
Questions? Contact the NRC Publications Archive team at
PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca. If you wish to email the authors directly, please see the first page of the publication for their contact information.
NRC Publications Archive
Archives des publications du CNRC
This publication could be one of several versions: author’s original, accepted manuscript or the publisher’s version. / La version de cette publication peut être l’une des suivantes : la version prépublication de l’auteur, la version acceptée du manuscrit ou la version de l’éditeur.
For the publisher’s version, please access the DOI link below./ Pour consulter la version de l’éditeur, utilisez le lien DOI ci-dessous.
https://doi.org/10.1039/b506695f
Access and use of this website and the material on it are subject to the Terms and Conditions set forth at
Ultra-trace determination of mercury in water by cold-vapor generation
isotope dilution mass spectrometry
Yang, L.; Willie, S. N.; Sturgeon, R.
https://publications-cnrc.canada.ca/fra/droits
L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.
NRC Publications Record / Notice d'Archives des publications de CNRC:
https://nrc-publications.canada.ca/eng/view/object/?id=b38065d7-597a-4620-95b9-95226aba6537
https://publications-cnrc.canada.ca/fra/voir/objet/?id=b38065d7-597a-4620-95b9-95226aba6537
Ultra-trace determination of mercury in water by cold-vapor
generation isotope dilution mass spectrometry
Lu Yang,* Scott Willie and Ralph E. Sturgeon
Institute for National Measurement Standards, National Research Council Canada, Ottawa, Ontario, Canada K1A 0R6. E-mail: Lu.Yang@nrc.ca
Received 11th May 2005, Accepted 19th July 2005
First published as an Advance Article on the web 11th August 2005
Determination of total Hg at the ppt level in water using cold vapor generation and inductively coupled plasma mass spectrometry for calibration by isotope dilution (ID) with both sector field (SF-ICP-MS) and time-of-flight MS (ICP-TOF-MS) detection is described. Samples were reduced on-line in a flow injection manifold with SnCl2to generate Hg vapor for both steady-state and transient sample introduction following
gold trapping. No significant difference in final Hg concentrations was detected using either approach. Certified reference material ORMS-3 river water (NRCC, Ottawa, Canada) was used for method validation. Mercury concentrations of 12.75 0.24 and 12.34 0.11 pg g1(mean and one standard deviation, n¼ 4) were obtained using steady-state generation with SF-ICP-MS and ICP-TOF-MS, respectively, in agreement with the certified value of 12.6 1.1 pg g1(combined uncertainty, k¼ 2). Concentrations of 12.55 0.41
and 12.65 0.30 pg g1(one standard deviation, n¼ 4) were obtained following gold trapping and SF-ICP-MS and ICP-TOF-SF-ICP-MS detection, respectively. Compared with SF-ICP-SF-ICP-MS, a 2-fold enhancement in the precision of the measured200Hg/201Hg and202Hg/201Hg ratios in a 100 ppt Hg standard solution was obtained using ICP-TOF-MS. Consequently, improved precisions of 0.87–2.33% RSD in Hg concentrations obtained by steady-state and transient signal acquisition, respectively, in ORMS-3 were obtained using ICP-TOF-MS, compared with 1.88–3.25% RSD obtained with SF-ICP-MS. Method detection limits (LODs, three times standard deviation) of 0.024 and 0.15 pg g1for Hg using direct CV and following gold trapping, respectively, were obtained using SF-ICP-MS. These were superior to corresponding LODs of 0.30 and 0.87 pg g1obtained using ICP-TOF-MS as a result of its 15-fold lower sensitivity.
Introduction
Mercury is a well-known hazard with exposure occurring principally through the consumption of water and fish, in which it bioaccumulates.1As a result, considerable effort has been devoted to the development of sensitive, accurate and rapid analytical methods for monitoring Hg in biological and environmental samples.
The low pg ml1concentrations of Hg typically found in natural waters present a significant challenge for its accurate quantitation.2,3Over the last decade, numerous methods have been developed for the determination of Hg in natural waters, snow and ice with reported detection limits in the range of 0.13 pg ml1(ppt) to 14 ng ml1(ppb).2–15Among these techniques, cold-vapor generation atomic absorption spectrometry (CVAAS) is one of the most commonly used owing to its simplicity, high sensitivity and sub-ppt detection power. Cold-vapor generation with/without preconcentration on a gold trap has also been coupled to inductively coupled plasma mass spectrometry (ICP-MS) for determination of Hg at ultratrace levels in waters.3–5,15Reported detection limits range from 0.13 to 0.20 ppt. Enhanced performance can be achieved when isotope dilution (ID) calibration is applied, which generally provides for higher accuracy and precision.16,17 Despite the advantages offered, however, few applications of ID for the determination of Hg in natural waters have been reported3,5,15 and quadrupole-based instruments (qICP-MS) have always been used.
The precision with which measurements can be made is directly influenced by the quality of the isotope ratio data. In the absence of severe mass bias corrections, matrix or spectral interferences, a limiting precision frequently arises which is
determined by the statistical arrival of ions at the detector, for which Poisson counting statistics apply to the intensities of the selected isotopes. Typical qICP-MS machines provide steady-state intensities of some 107counts s1ppm of analyte and are operated in a peak hopping mode, the duty factor of which for a given isotope is inversely related to the number of isotopes of interest. The precision with which isotope ratios can be mea-sured using sector field ICP-MS (SF-ICP-MS) is generally superior to that achieved by qICP-MS because its flat topped peaks produced in low resolution mode enable a more accurate and precise isotope ratio measurement.18 Additionally, these machines are capable of significantly increased sensitivity, some 109counts s1ppm being typical. The speed of current SF-ICP-MS in E-scan mode is competitive with qICP-MS,18,19 but the number of isotopes that can be simultaneously mea-sured remains limited when transient signal acquisition is required (e.g., vapor generation with gold trapping) if spectral skew is to be avoided. Time-of-flight based ICP-MS14,20–25 (ICP-TOF-MS) permits generation of a complete mass spec-trum from gated packets of ions, simultaneously eliminating correlated noise sources and the spectral skew inherent in scanning spectrometers. This should give rise to improved precision of isotope ratio measurements when multi-isotopic monitoring is performed on transient signals. Unfortunately, these instruments currently suffer from inferior sensitivity, typically 106counts s1ppm. When tasked with quantitation of ultratrace levels of mercury by ID (10 ppt and lower), an inherent Poisson limited precision in the isotope ratio (assum-ing an enriched isotopic spike is added such as to result in a 1 : 1 intensity ratio in the selected pair), and hence in the resultant concentration, of 4.5% RSD using a qICP-MS, 0.45% using a SF-ICP-MS and 14% with the ICP-TOF-MS arises, based on
A R T I C L E
www.rsc.org/jaas
DOI:
10.1039/b506695f
the above generic instrument sensitivities and assuming a 10 s acquisition time for steady-state sample introduction. These figures of merit can be improved by acquiring data for longer periods of time (assuming no limitation on sample volume) or by increasing the efficiency of sample introduction [using cold vapor (CV) generation3 or SPME sampling of vapor14]. CV generation is particularly attractive as it can be conveniently coupled to a gold amalgamation trap. In this manner, the Poisson limited RSD for ID ratio precision can be improved to 0.63%, 0.06% and 2%, respectively, for the three ICP-MS instruments based on a 50-fold improvement in sensitivity with CV sample introduction. These calculations suggest that cur-rent ICP-TOF-MS instrumentation is incapable of achieving the level of precision required for routine analysis of ultratrace concentrations of Hg. It is noteworthy that Evans et al.26
reported on a comparative study of the precision of isotopic ratio measurements for mercury using qICP-MS, SF-ICP-MS, ICP-TOF-MS and multi-collector-ICP-MS. Results obtained with the ICP-TOF-MS system were vastly inferior to those generated by the multi-collector but comparable to the perfor-mance of the qICP-MS. The concentration of mercury vapor admitted to each spectrometer was not stated but was sufficient to generate a steady-state signal by heating a sample of Almaden mercury ore. Thus, the relative performance char-acteristics of the TOF analyzer cannot be assessed from this study when applications relating to ID measurement of ultra-trace mercury are contemplated.
The objective of this study was thus to compare the relative performance of SF-ICP-MS and ICP-TOF-MS for the deter-mination of total Hg present at the ppt level in natural water using isotope dilution calibration with cold vapor generation. Both steady-state and transient (following trapping of gener-ated mercury on a gold foil) sample introduction were used. While this study does not provide a generic performance comparison of these two instrument designs for high precision Hg ratio measurements, a relative assessment of their use for the quantitation of ultratrace levels of mercury in real samples by ID is achieved. Reverse isotope dilution was performed to quantify an enriched201Hg concentration in a spike solution, thereby ensuring the quality of the final results. The method was validated by the determination of Hg in National Research Council Canada ORMS-3 water CRM.
Experimental section
Instrumentation
A ThermoFinnigan Element2 SF-ICP-MS instrument fitted with a plug-in quartz torch with a sapphire injector and a Ag guard electrode (Bremen, Germany) was used. Optimization of the Element2 was performed as recommended by the manu-facturer; typical operating conditions are summarized in Table 1. Isotopes of 200Hg, 201Hg and 202Hg were simultaneously monitored. For trapping of the mercury on gold and subse-quent transient data acquisition, the dwell time on each isotope was varied, as were the number of channels used to character-ize each mass peak so as to provide optimum precision in the Hg ratio measurements. Detector dead time was obtained using the approach of Nelms et al.27(method 2), wherein the
238
U/235U ratio was measured in standard solutions at con-centrations of 0.5, 1.0 and 2.5 ng ml1. A dead time of 18 1 ns (one calculated uncertainty) was derived from a plot of the measured238U/235U ratios versus U concentration at a slope of zero. Mass calibration was only performed once a week as a result of the stability of the instrument.
Analyses were also undertaken on an Optimass 8000 ortho-gonal acceleration (oa) ICP-TOF-MS instrument (GBC Scien-tific Equipment Pty. Ltd., Australia). Detailed information on the instrument and its performance characteristics have been reported earlier.20 Optimization of the Optimass 8000 was
performed as recommended by the manufacturer and typical conditions are summarized in Table 2.
A PerkinElmer (Norwalk, CT, USA) flow injection mercury system (FIMS) fitted with a gold trap was used for on-line cold vapor generation. The detection system on the FIMS was by-passed and, as shown in Fig. 1, the generated vapor was directed to the ICP torch. Tygon pump tubing was used to deliver sample and reagents: 1.14 mm id for the sample and 0.76 mm id for the reductant. All connections and tees were 0.25–28 low pressure Tefzel flangeless fittings (Upchurch Scien-tific Inc., Oak Harbor, WA, USA) and 1.58 mm od PTFE tubing (Cole Parmer, Illinois, USA). A gas–liquid separator (10 ml volume, PerkinElmer, part B0507959) ensured efficient stripping of the generated vapor from the liquid phase. An additional subsequent liquid trap ensured that no large liquid droplets nor acid vapor were transported to the ICP torch. Coupling of the FIMS to the mass spectrometers was accom-plished by directing generated vapor to the torch through a 0.5 m length of Teflon lined Tygon tubing (0.2500 od). A
polypropylene 0.2500 Swagelok T connected the instrument nebulizer gas to the FIMS Ar carrier gas. The two gas flows entered the torch through a ball joint adapter. Following a daily optimization of the spectrometers using liquid sample introduction, the plasma was then extinguished and the spray chamber and nebulizer assembly replaced with the transfer line. Final optimization of the lens voltages and nebulizer gas flow for dry plasma conditions was performed by monitoring Table 1 SF-ICP-MS operating conditions
Rf power 1150 W Plasma Ar gas flow rate 15.0 l min1 Auxiliary Ar gas flow rate 1.05 l min1 Ar carrier gas flow rate 0.69 l min1 Sampler cone (nickel) 1.1 mm Skimmer cone (nickel) 0.8 mm
Lens voltage Focus:830 V; x deflection: 0.95 V; y deflection: 1.60 V; shape: 105 V
Resolution 300
Data acquisition Steady-state: E-scan, 5 runs 70 passes, 5% mass window, 5 ms sample time, 200 samples per peak for direct CV Transient: E-scan, 400 runs, 5% mass window, 5 ms sample time, 200 samples per peak
Table 2 ICP-TOF-MS operating conditions Rf power 1150 W Plasma Ar gas flow rate 10 l min1
Auxiliary Ar gas flow rate 1.00 l min1
Ar carrier gas flow rate 0.60 l min1 Sampler cone (stainless steel) 1.1 mm Skimmer cone 1 (stainless steel) 1.0 mm Skimmer cone 2 (stainless steel) 1.0 mm
Voltage Skimmer:950 V; extraction: 1360 V; Z1: 950 V; Y mean: 300 V; Y deflection: 0 V; Zlens mean:920 V; Zlens deflection: 0 V; lens body:145 V; multiplier gain: 2400 V
Data acquisition 55 s acquisition time for direct CV; 100 spectra collected over 4 s for CV with trapping on gold
J . A n a l . A t . S p e c t r o m . , 2 0 0 5 , 2 0 , 1 2 2 6 – 1 2 3 1 1227
Published on 11 August 2005. Downloaded by National Research Council Canada on 26/11/2015 16:11:42.
the Hg signal arising from continuous CV generation (by-passing the gold trap) of a 100 ppt Hg standard solution. The nebulizer flow consisted of a combination of the mass spectrometer setting (typically 0.65 l min1) and the Ar carrier flow from the FIMS (preset to 0.30 l min1).
Reagents and solutions
Hydrochloric acid was purified in-house by sub-boiling dis-tillation of reagent grade feedstock in a quartz still prior to use. High purity de-ionized water (DIW) was obtained using a NanoPure mixed bed ion exchange system fed with reverse osmosis domestic feed water (Barnstead/Thermolyne Corp, Iowa, USA). A 0.5% solution of SnCl2 was prepared by
dissolving stannous chloride (Fisher Scientific, Nepean, Cana-da) in 1% HCl and purging with high purity argon (99.99%, BOC Gases, Mississauga, Ontario, Canada) for at least 2 hours prior to use as the reductant in order to minimize contamina-tion.
‘‘Natural abundance’’ high purity Hg metal was purchased from Johnson Matthey & Co. Ltd (London, UK). A 1000 mg g1 stock solution was prepared in 5% HCl. A working standard (101 ng g1) used for201Hg reverse isotope dilution, was prepared by serial dilution of the stock in 2% HCl.
Enriched 201Hg isotope was purchased from Oak Ridge National Laboratory (USA) as the metal. A 300 mg g1stock solution was prepared in 5% HCl. A spike solution of 1.9 ng g1 was prepared by serial dilution in 2% HCl. Accurate concentration of the enriched spike was determined by reverse isotope dilution using the gravimetrically prepared natural abundance Hg standard.
The CRM ORMS-3 (Mercury in River Water Reference Material, NRCC, Ottawa, Canada) was used for validation of methodology. This CRM is packaged glass ampoules.
Sample preparation and analysis procedure
Sample preparation was conducted in a class-100 clean room. Freshly spiked samples of ORMS-3 were prepared daily to minimize any possible contamination from the laboratory environment. Each of four replicate vials of ORMS-3 (each containing about 50 g) was opened and 0.100 g of 1.9 ng g1 enriched201Hg spike was added directly to the vial. Vials were then covered with plastic wrap and vortex shaken prior to analysis. Sample mass was obtained by the difference in weight between the opened vial of ORMS-3 and the dried empty vial after the analysis. A 500 g solution of 2% HCl spiked with the same mass of201Hg used for ORMS-3 was used to estimate method detection limits.
Four reverse-spike isotope dilution calibration samples were prepared to quantify the concentration of the enriched201Hg spike. A 1.00 g aliquot of 1.9 ng g1 enriched 201Hg spike
solution and 0.100 g of 101 ng g1 natural abundance Hg solution were accurately weighed into a pre-cleaned high density polyethylene bottle and diluted with 50 g of 2% HCl. For direct CV analysis, both FIMS pumps were operated at 100 rpm to provide for sample and reductant flows of 10 and 4 ml min1, respectively. The main valve was set to the fill position and an Ar flow of 0.3 l min1 transported the generated Hg vapor through the gas–liquid separator and acid wash trap to the plasma. Following 70 s of sample uptake, data acquisition was initiated and collected for an additional 55 s. Isotopes of 200Hg, 201Hg and 202Hg were simultaneously monitored and200Hg/201Hg and202Hg/201Hg ratios were
cal-culated. The Hg concentration was calculated from the average of both 200Hg/201Hg and 202Hg/201Hg ratios. In this way, compensation for mass bias was accomplished without resort-ing to an external mass bias correction as a result of its opposite influence on these two isotope pairs. This approach has been successfully applied to the determination of butyltins by Rodrı´guez-Gonza´lez et al.28 using two pairs of Sn ratios with calibration by ID.
The sequence of operating steps for the FIMS amalgamation system is summarized in Table 3. During step 1, the sample tubing leading up to the valve is flushed with new sample solution. In step 2, the valve is switched to the ‘fill’ position to permit the sample to be mixed on-line with 0.5% SnCl2. The
generated vapor is transported in a 0.3 l min1 flow of Ar through a gas–liquid separator and an acid trap and subse-quently accumulated on the Pt–Au foil. Following a 60 s sample loading at a flow rate of 10 ml min1, the valve is switched to the injection position and pump 2 is operated for 30 s to ensure that all Hg0was swept through the system to the
Pt–Au foil (step 3). Heat is applied to the foil and Hg0 is released to the ICP for measurement in step 4. Finally, a flow of air is directed at the outer surface of the trap to cool the Pt/ Au foil prior to the next run. Runs consisting of 0.5% SnCl2in
1% HCl were repeated through at least 5 cycles (for cleaning the entire on-line FIMS amalgam system) before processing the real samples. The volume of sample and blank processed was 10 ml. Data acquisition on both spectrometers was manually triggered after an appropriate delay time (80 s) to sample the transient peaks. The acquired data were transferred to an off-line computer for further processing using in-house software to yield peak areas to generate 200Hg/201Hg and 202Hg/201Hg ratios, from which the analyte concentration in the sample was calculated.
Results and discussion
Isotope ratio measurements for Hg
The performance of the SF-ICP-MS and ICP-TOF-MS instru-ments for measureinstru-ments of the200Hg/201Hg and202Hg/201Hg ratios used for typical ID calibration purposes (i.e., a 1 : 1 ratio of endogenous 202 to spiked 201) were compared. A 100 ppt solution of Hg was used for CV generation to provide both steady-state and transient signals (following trapping on gold). Fig. 1 Schematic diagram of the FIMS/ICP-MS system.
Table 3 FIMS program for trapping Hga
Step Time/s Pump 1/rpm
Pump
2/rpm Valve Cool Argon Function 1 15 120 0 Injection On On Pre-fill 2 60 100 100 Fill Off On Load 3 30 0 100 Injection Off On Rinse 4 10 0 0 Injection Off On Heat 5 20 0 0 Injection On On Cooling
aA sample volume of 10 ml was processed with each ‘‘injection’’.
In contrast to pneumatic sample introduction of the liquid solution, signals derived from CV generation suffered substan-tially increased noise (14% RSD) arising from both the reduction and the gas–liquid separation process, as a segmen-ted flow gas stripping arrangement was used. Consequently, the precision of measured Hg ratios based on individual points is also affected. Thus,200Hg/201Hg and202Hg/201Hg ratios were
calculated based on the summed counts over the full integra-tion (55 s measurement) period. Results are summarized in Table 4.
Although the total integration time was maintained identical for both instruments, acquisition conditions were different for the two MS configurations. Over a 55 s read period, the ICP-TOF-MS summed detector response from 1.76 106 (55
29 411) oa pushouts. As the signal intensity was outside the linear range of pulse counting for the TOF, analog detection was used. This provided a typical signal of 3.4 105cps for 202
Hg from a 100 ppt Hg standard with continuous CV generation. Cross-calibration of the detector yielded an amplification factor of 35 when changing from pulse counting to analog output, with the result that the equivalent ion count rate (for the purposes of estimating counting statistics) was 9.7 103
cps. For the SF-ICP-MS, 5 runs of 70 passes with a 5 ms dwell time on each of 10 channels within a 5% mass defining the isotope peak resulted in a typical signal of 7.0 106
cps for 202Hg under these operating conditions. With a 55 s integration time, the total counts recorded for202Hg was 5.3 105
with the TOF and 3.85 108 with the SF instrument. Poisson statistics applied to these data provide a count rate limited precision achievable for the202Hg/201Hg ratio of 0.19% for the TOF and 0.007% with the SF. These values compare well with the experimental precision (Table 4) of 0.26% with the TOF but that for the SF is degraded to 0.31% RSD. Clearly, the compensation for correlated noise, which is high with this sample introduction system, is achieved with the simultaneous measurement cap-abilities of the TOF, but remains a significant noise component for the SF because of the sequential nature of the data acquisition.
A significant enhancement in the Hg signal (65-fold in peak sensitivity) could be obtained following trapping of all of the mercury generated from a 10 ml sample onto the Pt–Au foil. Desorption transients were similar in shape to those reported by Evans et al.26and were characterized by a full width at half-maximum of 1s and a base width of 3 s, as is evident from Fig. 2. Mercury isotope ratios derived from transient signals were based on the integrated counts over the full transient. Based on the results of processing 10 ml volumes of the 100 ppt standard, the isotope ratio precisions for202Hg/201Hg were characterized
by a 0.53% RSD with the SF- and 0.44% with the TOF-MS (Table 4). As expected, the resulting precision of the calculated ratios obtained with steady-state CV generation is some 2-fold better than that obtained with transient sample introduction for both detection systems, an observation frequently reported in the literature.29It is not possible to compare these data with those reported by Evans et al.26 because the total flux of mercury sampled for transient measurements by these authors was unknown.
Determination of Hg in ORMS-3 using ID with continuous CV generation
The quality of data obtained with ID calibration is largely dependent on the accuracy and precision of the ratio measure-ments of the selected isotope pairs. An important aspect relating to the accuracy of ratio measurements is the correction undertaken for mass discrimination or mass bias. This can be achieved using an external correction based on the expected ratios divided by the measured ratios from a natural abun-dance Hg standard having known isotopic composition. Rod-rı´guez-Gonza´lez et al.28recently reported a new approach to mass bias correction (internal mass bias correction) which obviated the need to run a natural abundance standard. The approach was based on use of the average concentration derived from two pairs of ratios (e.g.,118Sn/119Sn and 120Sn/
119Sn in their study) whose mass bias correction effects
oc-curred in opposite directions. The derived average concentra-tion was thus free of mass bias effects. This approach not only saves analysis time, but also reduces the uncertainty associated with such measurements since mass bias correction is elimi-nated. For this study, both200Hg/201Hg and202Hg/201Hg ratios were measured and this internal mass bias correction approach was used for the determination of Hg in ORMS-3 water.
The intensities of the Hg isotopes measured during intro-duction of only the SnCl2 reductant (sample introduction
Table 4 Isotope ratios with SF-ICP-MS and ICP-TOF-MS
Methods Measured ratios (one standard deviation n¼ 6)
200Hg/201Hg 202Hg/201Hg SF-ICP-MS CVa 1.7511 0.0066 2.2696 0.0071 CV with Au trappingb 1.750 0.011 2.269 0.012 ICP-TOF-MS CVa 1.7512 0.0031 2.2687 0.0058 CV with Au trappingb 1.7529 0.0070 2.265 0.010 a
55 s integration period, 100 ppt standard.b10 ml sample volume, 100 ppt standard.
Fig. 2 Chromatograms characterizing a201Hg spiked ORMS-3
ob-tained using CV with trapping on Au. (a) SF-ICP-MS detection; (b) ICP-TOF-MS detection. Black trace:202Hg; grey trace:201Hg; and
closed circles:202Hg/201Hg ratio.
J . A n a l . A t . S p e c t r o m . , 2 0 0 5 , 2 0 , 1 2 2 6 – 1 2 3 1 1229
Published on 11 August 2005. Downloaded by National Research Council Canada on 26/11/2015 16:11:42.
stopped by pinching the uptake line) served to establish the measurement baseline from which data generated with all spiked ORMS-3 samples were subtracted to yield a blank corrected response. Similarly, the intensities of the Hg isotopes measured in 2% HCl were subtracted from all reverse ID calibration solutions. The following equation was used for quantitation of Hg in ORMS-3 using ID with either SF-ICP-MS or ICP-TOF-SF-ICP-MS: Cx¼ Cz my mx mz m0 y Ay By Rn Bx Rn Ax Bz R 0 n Az Ay By R0n AWx AWz ð1Þ
where Cx is the analyte concentration in pg g1; Cz is the
concentration of natural abundance Hg standard (pg g1); my
is the mass (g) of spike used to prepare the blend solution of sample and spike; mxis the mass (g) of sample used; mzis the
mass (g) of natural abundance Hg standard used; m0
y is the
mass (g) of spike used to prepare the blend solution of spike and natural abundance Hg standard solution; Ayis the
abun-dance of the reference isotope (200Hg or202Hg) in the spike; B y
is the abundance of the spike isotope (201Hg) in the spike; Ax
(or Az) is the abundance of the reference isotope (200Hg or 202Hg) in the sample (or in the standard); B
x (or Bz) is the
abundance of the spike isotope (201Hg) in the sample (or in the standard); Rnis the measured reference/spike isotope ratio in
the blend solution of sample and spike; R0
nis the measured
reference/spike isotope ratio in the blend solution of spike and natural abundance Hg standard; AWx is the analyte atomic
weight in the sample and AWzis the analyte atomic weight in
the natural abundance Hg standard.
A concentration of 12.75 0.24 pg g1was calculated for
Hg (one standard deviation, n ¼ 4), based on the average values obtained from the200Hg/201Hg and202Hg/201Hg ratios
using continuous sample introduction, CV generation and SF-ICP-MS detection. This result is in good agreement with the certified value of 12.6 1.1 pg g1(combined uncertainty, k¼ 2). Using continuous CV generation with ICP-TOF-MS detec-tion, a concentration of 12.34 0.11 pg g1 (one standard
deviation, n¼ 4) was obtained.
The analytical blank for steady-state CV generation was defined by the baseline noise generated by the acidified reduc-tant. Method detection limits (LODs based on three times the standard deviation of the concentration of the blank) for ID-CV-SF-ICP-MS and ICP-TOF-MS techniques were calculated based on eight measurements of a201Hg spiked sample blank. Values of 0.024 and 0.30 pg g1were obtained using SF-ICP-MS and ICP-TOF-SF-ICP-MS, respectively. The superior LOD ob-tained with SF-ICP-MS is a direct consequence of its enhanced sensitivity compared with ICP-TOF-MS.
Determination of Hg in ORMS-3 using ID with purge and trap CV generation
Trapping mercury vapor on Au is frequently used to achieve preconcentration in order to improve sensitivity at low con-centrations.4,26 The Pt–Au foil used in these experiments suffered from memory effects in that a carryover signal was generated following trapping of analyte from the sample. This may have been a consequence of the limited heating capabil-ities of the FIMS system which relies on radiant heat from several IR lamp sources. This problem should be eliminated by use of gold coated sand for trapping,26wherein the gold layer is very thin (mm). Five heating cycles of the trap were thus necessary to reduce the signal to background levels following the processing of a 100 ppt standard solution. Consequently, five blank samples were required to be run before the Hg signal returned to the original baseline following the analysis of a 100 ppt standard solution.
As is evident in Fig. 2a, spectral skew arises for the transient signals recorded with the SF-ICP-MS as a consequence of the
sequential nature of the data acquisition; measurements made with the ICP-TOF-MS are free from this aberration. As a consequence, 202Hg/201Hg ratios for spiked ORMS-3 calcu-lated at intervals across the transient signals using the SF display a continuous decrease with time (Fig. 2a). This may be due to isotope fractionation during thermal release of Hg from the gold trap, as noted by Evans et al.,26 but data provided by the TOF detection system for the same transient sample introduction do not support this. Fig. 2b illustrates that relatively constant isotope ratios can be obtained across tran-sient signals recorded with the TOF, subject to the limitations of counting statistics. Isotope ratios used for quantitation were thus based solely on the integrated response for each isotope transient, thereby providing optimum precision of measure-ment.26
No significant difference was evident in the ID results for ORMS-3 using either SF- or TOF-MS detection; concentra-tions of 12.55 0.41 and 12.65 0.30 pg g1(one standard deviation, n¼ 4) were obtained, respectively. The reproduci-bility (3.3 and 2.3% RSD) of these results is poorer than that achieved using CV generation with steady-state sample intro-duction (1.9 and 0.87% RSD).
Method detection limits based on measurement of transient signals were estimated from replicate measurements of the blank (1.4 pg g1), yielding values of 0.15 and 0.87 pg g1 for Hg by SF- and TOF-MS, respectively. These LODs are inferior to those obtained with direct CV generation due to the inherently lower precision typically characterizing transient signals.29
Conclusion
Quantitation of total Hg by CV generation ID at ultra-trace concentrations in water can be achieved using ICP-TOF-MS detection with high precision and accuracy and a LOD of 0.9 pg g1. Jitaru and Adams14reported a detection limit of 0.27 pg ml1for Hg using solid phase microextraction coupled to GC-MS. The often expressed belief that ICP-TOF-MS is incapable of such performance because of its reduced sensitivity relative to even qICP-MS is thus fallacious. Mea-surement performance is competitive with the more expensive SF-ICP-MS instrumentation if used for such determinations and precision of ratio measurements obeys Poisson statistics. Barbaste et al.30reported similar performance comparisons for the determination of Pb isotope ratios in wine when ICP-TOF-MS, qICP-MS and multi-collector-ICP-MS machines were used for detection.
References
1 P. J. Craig, Organometallic Compounds in the Environment: Prin-ciples and Reactions, Longman, Harlow, 1986.
2 J. Allibone, E. Fatemian and P. J. Walker, J. Anal. At. Spectrom., 1999, 14, 235–239.
3 J. L. Mann, S. E. Long and W. R. Kelly, J. Anal. At. Spectrom., 2003, 18, 1293–1296.
4 E. Debrah, E. R. Denoyer and J. F. Tyson, J. Anal. At. Spectrom., 1996, 11, 127–132.
5 S. Willie, Am. Lab., 2002, 34, 40–43.
6 L. Manganiello, A. Rı´os and M. Valca´rcel, Anal. Chem., 2002, 74, 921–925.
7 J. C. A. de Wuilloud, R. G. Wuilloud, M. F. Silva, R. A. Olsina and L. D. Martinez, Spectrochim. Acta, Part B, 2002, 57, 365–374. 8 J. L. Manzoori, M. H. Sorouraddin and A. M. Haji Shabani,
J. Anal. At. Spectrom., 1998, 13, 305–308.
9 J. L. Go´mez-Ariza, F. Lorenzo and T. Garcı´a-Barrera, J. Chro-matogr. A, 2004, 1056, 139–144.
10 B. San Vicente de la Riva, J. M. Costa-Ferna´ndez, W. J. Jin, R. Pereiro and A. Sanz-Medel, Anal. Chim. Acta, 2002, 455, 179–186. 11 P. F. Schuster, D. P. Krabbenhoft, D. L. Naftz, L. D. Cecil, M. L. Olson, J. F. Dewild, D. D. Susong, J. R. Green and M. L. Abbott, Environ. Sci. Technol., 2002, 36, 2303–2310.
12 C. F. Boutron, G. M. Vandal, W. F. Fitzgerald and C. P. Ferrari, Geophys. Res. Lett., 1998, 25, 3315–3318.
13 C. P. Ferrari, A. L. Moreau and C. F. Boutron, Fresenius’ J. Anal. Chem., 2000, 366, 433–437.
14 P. Jitaru and F. C. Adams, J. Chromatogr. A, 2004, 1055, 197–207. 15 R. G. Smith, Anal. Chem., 1993, 65, 2485–2488.
16 P. De Bie`vre in Isotope Dilution Mass Spectrometry in Trace Element Analysis in Biological Specimens, eds. R. F. M. Herber and M. Stoeppler, Elsevier, 1994.
17 K. G. Heumann in Isotope Dilution Mass Spectrometry in Inor-ganic Mass Spectrometry, eds. F. Adams, R. Gijbels and R. Van Grieken, John Wiley & Sons, 1988.
18 L. Yang, Z. Mester and R. E. Sturgeon, J. Anal. At. Spectrom., 2003, 18, 1365–1370.
19 ELEMENT2 Thermo Finnigan Application Flash Report No. E15, ThermoFinnigan, November, 2002.
20 R. E. Sturgeon, J. W. Lam and A. Saint, J. Anal. At. Spectrom., 2000, 15, 607–616.
21 X. Tian, H. Emteborg and F. C. Adams, J. Anal. At. Spectrom., 1999, 14, 1807–1814.
22 F. Vanhaecke, L. Moens, R. Dams, L. Allen and S. Georgitis, Anal. Chem., 1999, 71, 3297–3303.
23 H. Emteborg, X. Tian, M. Ostermann, M. Berglund and F. C. Adams, J. Anal. At. Spectrom., 2000, 15, 239–246.
24 S. N. Willie and R. E. Sturgeon, Spectrochim. Acta, Part B, 2001, 56, 1707–1716.
25 K. Benkhedda, H. Goenaga Infante, E. Ivanova and F. C. Adams, J. Anal. At. Spectrom., 2000, 15, 1349–1356.
26 R. D. Evans, H. Hintelmann and P. J. Dillon, J. Anal. At. Spectrom., 2001, 16, 1064–1069.
27 S. M. Nelms, C. R. Que´tel, T. Prohaska, J. Vogl and P. D. P. Taylor, J. Anal. At. Spectrom., 2001, 16, 333–338.
28 P. Rodrı´guez-Gonza´lez, J. Ruiz Encinar, J. I. Garcı´a Alonso and A. Sanz-Medel, J. Anal. At. Spectrom., 2002, 17, 824–830. 29 I. Gu¨nther-Leopold, B. Wernli, Z. Kopajtic and D. Gu¨nther,
Anal. Bioanal. Chem., 2004, 378, 241–249.
30 M. Barbaste, L. Halicz, A. Galy, B. Medina, H. Emteborg, F. C. Adams and R. Lobinski, Talanta, 2001, 54, 307–317.
J . A n a l . A t . S p e c t r o m . , 2 0 0 5 , 2 0 , 1 2 2 6 – 1 2 3 1 1231
Published on 11 August 2005. Downloaded by National Research Council Canada on 26/11/2015 16:11:42.
